We are facing an epidemic of indeterminate pulmonary nodules (IPNs), not only those found incidentally, but also through the proliferation of CT screening programs targeting high-risk individuals for lung cancer following the encouraging results of the National Lung Screening Trial (NLST). Although the large majority of IPNs are benign, there are currently no effective predictive tools to discriminate benign from malignant nodules, leading to a large number of follow-up CTs, unnecessary invasive biopsies, anxiety, and wasted healthcare spending. Our project addresses this deficiency in knowledge, which we expect will result in improved non-invasive testing and paradigm-shifting prediction models in the clinical management of patients with IPNs. This project has the potential to transform the management of individuals presenting with IPNs. We aim at obtaining significant gains in non-invasive diagnostic accuracy of screening-detected and incidentally-detected IPNs by integrating clinical variables, innovative quantitative structural CT variables with novel blood-based biomarker candidates of IPNs. We will produce actionable diagnostic prediction models to personalize care and identify the small subset of individuals who will need a definitive intervention for lung cancer, and conversely, others who are at very low risk and do not require extensive follow-up. Based on findings from this project, we anticipate 1) using longitudinal data to improve our prediction accuracy for IPNs over time; 2) reduction in the rate of unneeded thoracotomies, the burden of unnecessary follow-up including anxiety; 3) ultimately increase the rate of cure and reduction of unnecessary radiation and healthcare cost.